Forced Convection Heat Transfer Enhancement by Porous Pin Fins in Rectangular Channels

Transcription

1 Jian Yang Min Zeng Qiuwang Wang 1 wangqw@mail.xjtu.edu.cn State Key Laboratory of Multiphase Flow in Power Engineering, School of Energy and Power Engineering, Xi an Jiaotong University, Xi an, Shaanxi , China Akira Nakayama Department of Mechanical Engineering, Shizuoka University, Johoku, Hamamatsu , Japan Forced Convection Heat Transfer Enhancement by Porous Pin Fins in Rectangular Channels The forced convective heat transfer in three-dimensional porous pin fin channels is numerically studied in this paper. The Forchheimer Brinkman extended Darcy model and two-equation energy model are adopted to describe the flow and heat transfer in porous media. Air and water are employed as the cold fluids and the effects of Reynolds number (Re), pore density (PPI) and pin fin form are studied in detail. The results show that, with proper selection of physical parameters, significant heat transfer enhancements and pressure drop reductions can be achieved simultaneously with porous pin fins and the overall heat transfer performances in porous pin fin channels are much better than those in traditional solid pin fin channels. The effects of pore density are significant. As PPI increases, the pressure drops and heat fluxes in porous pin fin channels increase while the overall heat transfer efficiencies decrease and the maximal overall heat transfer efficiencies are obtained at PPI 20 for both air and water cases. Furthermore, the effects of pin fin form are also remarkable. With the same physical parameters, the overall heat transfer efficiencies in the long elliptic porous pin fin channels are the highest while they are the lowest in the short elliptic porous pin fin channels. DOI: / Keywords: porous pin fin channel, forced convection, heat transfer enhancement, CFD simulation 1 Introduction Pin fins have a variety of applications in industry due to their excellent heat transfer performance, e.g., in cooling of electronic components, in cooling of gas turbine blades, and recently, in hot water boilers of central heating systems, etc. 1. In the two early studies by Sahiti et al. 2,3, it was demonstrated that pin fin arrays offer the most effective way of enhancing the heat transfer rate within a particular heat exchanger volume. However, the pressure drops in such heat exchangers are usually much higher than those in others 1 ; this defect greatly lowers the overall heat transfer performances of pin fin heat exchangers and as a result, their applications are restricted. In order to reduce the pressure drops and improve the overall heat transfer performances for pin fin heat exchangers, porous metal pin fin arrays may be used instead of traditional solid metal pin fin arrays. As porous media can significantly intensify the mixing of fluid flow and increase the contact surface area with fluid inside, it has been regarded as an effective way to enhance heat transfer by using porous media 4. The flow and heat transfer in porous pin fin heat exchangers for present study can be modeled as forced convective heat transfer in partially filled porous channels. The researches on forced convection with partially filled porous configurations have been investigated extensively in the last years. Hadim 5 studied the laminar forced convection in a fully or partially filled porous channel containing discrete heat sources on the bottom wall. The Brinkman Forchheimer extended Darcy model were used for the computations. He found that when the width of the heat source and the space between the porous layers were of same magnitudes as the channel height, the heat transfer enhancement in the partially filled channel was almost the same as that in the fully filled porous channel while the pressure drop was 1 Corresponding author. Contributed by the Heat Transfer Division of ASME for publication in the JOUR- NAL OF HEAT TRANSFER. Manuscript received January 8, 2009; final manuscript received November 7, 2009; published online March 5, Assoc. Editor: S. A. Sherif. much lower. Hadim and Bethancourt 6 later studied the similar problem in a partially filled porous channel. They found that when the heat source width was decreased, there was a moderate increase in heat transfer enhancement and a significant decrease in pressure drop. Huang and Vafai 7 presented a detailed investigation of forced convection in a channel filled with multiple emplaced porous blocks. With comparison of the local Nusselt number distributions between the channel with and without porous blocks, they found that significant heat transfer augmentation can be achieved through the emplacement of porous blocks. Huang et al. 8 later presented a similar investigation in cooling of multiple heated blocks covered with porous media. The results showed that significant cooling augmentation of the blocks can be achieved through the cover of finite-sized porous substance. Other similar studies of forced convection in a channel filled with porous blocks can also be found in Refs. 9,10. Besides porous block, porous baffles are also popular for heat transfer enhancement applications. Ko and Anand 11 experimentally studied the heat transfer enhancement in a rectangular channel by using a porous baffle made up of aluminum foam. The experiments showed that the use of porous baffles resulted in heat transfer enhancement as high as 300% compared with heat transfer in straight channel with no baffles and the heat transfer enhancement ratio was found to be higher for taller and thicker porous baffles. Furthermore, Yang and Huang 12 presented a numerical prediction on the turbulent fluid flow and heat transfer characteristics for rectangular channel with porous baffles. They found that, both the solid and porous baffles walls enhanced the heat transfer relative to the smooth channel while the porous baffle channel has a lower friction factor due to less channel blockage. According to above studies, it can be concluded that with proper selection of governing parameters, significant heat transfer augmentation and pressure drop reduction can be achieved simultaneously in partially filled porous channels. Therefore, we could believe that the overall heat transfer performance of porous pin fin heat exchangers with proper configurations would be better than that of traditional solid pin fin heat exchangers. On account of this Journal of Heat Transfer Copyright 2010 by ASME MAY 2010, Vol. 132 /

2 Fig. 2 Different forms of porous pin fin cross-section: a circular form, b cubic form, c long elliptic form, and d short elliptic form Fig. 1 Physical model: a porous pin fin heat sink and b representative computational domain reason, in our previous study as reported by Yang et al. 13, we performed a comprehensive numerical study on forced convection heat transfer in three-dimensional 3D porous pin fin channels with air as the cold fluid. We found that with the proper selection of governing parameters, the pressure drops in porous pin fin channels were much lower than those in solid pin fin channels while the heat fluxes and the overall heat transfer efficiencies were much higher. The overall heat transfer efficiencies in long elliptic porous pin fin channels were the best and the maximal values were obtained at K= m 2. These findings could be useful for understanding and optimizing the flow and heat transfer performances in porous pin fin heat exchangers. However, it was noted that, in our previous study 13, the permeability K and inertial coefficient c F in the momentum equation were modeled with Ergun equation 14 and the volumetric heat transfer coefficient h v in the energy equation was calculated with Wakao equation 15. This would be reasonable for the flow and heat transfer in the packed beds of particles while for the porous media with high porosity =0.9, such as metal foams, the applicability of the Ergun equation and Wakao equation would be questionable. Furthermore, in the work of Yang et al. 13, only air was investigated and the performances for other fluids are still unknown, which would also be important for applications. With these motivations in the present study, we further study the forced convection heat transfer in three-dimensional porous pin fin channels and the performances for both air and water are carefully compared. According to the authors knowledge, almost no such attentions have been paid on this subject before. The Forchheimer Brinkman extended Darcy model and two-equation energy model with more reasonable model parameters K, c F, and h v for porous media are employed and the effects of Reynolds number, pore density, and pin fin form are studied in detail. 2 Physical Model and Computational Method 2.1 Physical Model and Dimensions. As shown in Fig. 1 a, the physical model is derived from traditional pin fin heat sink, which generally consists of a bottom wall, two side walls, a top wall, and a pin fin array. The bottom wall is hot and its temperature is kept at T h. The side and the top walls are kept adiabatic. The pin fin array is made of high porosity metal foams aluminum and arranged in stagger; air and water are used as the cold fluids. In order to obtain a basic understanding of flow and heat transfer performances in porous pin fin heat exchangers, a simplified porous pin fin channel with appropriate boundary conditions is adopted for the computations, which can be regarded as forced convection heat transfer in a partially filled porous channel Hadim 5, Huang and Vafai 7, and Yang et al. 13. The computational domain is depicted in Fig. 1 b, which is composed of a developing inlet block L 1 =10 mm, two pin fin array unit cells L 2 = mm, and a developing outlet block L 3 =70 mm. The dimensions of the computational domain are L mm W 3.26 mm H 10 mm for air and L mm W 3.26 mm H 2 mm for water, where the channel height H for water is much lower due to its higher heat transfer capacity. The total area of pin fin cross-sections over the base wall area in single pin fin array unit cell is 15%, which is reasonable for industry applications. The temperature and velocity of inlet are kept at T in and u in, respectively. The bottom wall of pin fin array unit cells is the hot wall and the temperature is kept at T h. Two other bottom walls and all top walls are kept adiabatic. The symmetry boundary conditions are adopted for two side walls and the flow and heat transfer of outlet are considered to be fully developed. Furthermore, four different kinds of porous pin fins with circular, cubic, long elliptic, and short elliptic cross-section forms are employed to investigate the pin fin configuration effects and the cross-section areas of different pin fins are identical with each other A pin =3.14 mm 2. The physical dimensions and cross-section forms of different porous pin fins are presented in Fig Governing Equations and Computational Method. The flow in the computational domain is considered to be threedimensional, laminar, incompressible, and steady for both clear fluid and porous regions. For clear fluid region, the flow and heat transfer are modeled with Navier Stokes and energy equations. For porous region, the metal foams are assumed to be homogeneous, isotropic, high porosity =0.9, and high thermal conductivity aluminum: k s =238 W m 1 K 1. The Forchheimer Brinkman extended Darcy model 16 is adopted to simulate the flow in porous media, where the inertia and viscosity effects are considered. Furthermore, the porous matrix is assumed to be in local thermal nonequilibrium with fluid phase inside due to their large thermal conductivity difference. Therefore, the two-equation energy model 16 is employed to account for the heat transfer between porous matrix and fluid inside. The conservation equations for mass, momentum, and energy are as follows. Continuity V = / Vol. 132, MAY 2010 Transactions of the ASME

3 Momentum Energy Clear fluid region: V V = 1 Porous region: 1 f p + v f 2 V 2 V V = 1 f p + v f 2 V v f K V c F K V V Clear fluid region: cp f V Tf = kf Tf Porous region: fluid phase: c 3 p f V T f = k f T f + h T s T f porous matrix: 0 = 1- k s T s + h v T f -T s 2 where V is the velocity vector. T f and T s are the temperatures of the fluid and porous matrix, respectively. is the porosity. The permeability K, Forchheimer coefficient c F, and volumetric heat transfer coefficient h v are calculated with following correlations developed from high porosity metal foams by Bhattacharya et al. 17 and Calmidi and Mahajan 18. K = d 2 p c F = G G 4 h v = 3 d fg 0.59d p 2 kf 0.52Pr 0.37 V d f / f 0.5 d f where Pr is the Prandtl number with the definition of Pr = f / k/ c p f. f is the kinetic viscosity of fluid and k f is the thermal conductivity of fluid. d p and d f are the pore size and fiber diameter of the metal foams, respectively. is the tortuosity of the porous matrix and G is a shape function that takes into account the variation in fiber cross-section with porosity. The definitions of d p, d f,, and G are as follows: d p = /PPI; d f = d p 3 G ; = G =1 exp 1 / G where PPI is the pore density of the metal foams. 1 ; 5 Boundary conditions x =0 Tf = Tin, u = uin, v = w =0 T f x = L x =0, u x = v x = w x =0 T f y =0 y =0 0 x L 1,L 2 x L, T f = T h L 1 x L 2 T s = T h porous region, u = v = w =0 T f y = H y =0, T s =0 porous region, u = v = w =0 y z =0,W T f z =0, T s u =0 porous region, z z = v =0,w =0 z The thermal physical quantities of interest in present investigation are the heat flux of the hot wall q, the pressure drop p, the overall heat transfer efficiency, and the heat transfer performance ratio, which are defined as follows: q = c p f u in A in T out T in A h ; p = p in p out ; = q p ; = Nu hav,p/nu hav,s f p /f s 1/3 where A in is the area of inlet. A h is the base area of hot wall and T out is the average temperature of outlet. Nu hav is the average Nusselt number of the hot wall. f is the friction factor. The subscripts p and s represent values obtained in porous and solid 6 Journal of Heat Transfer MAY 2010, Vol. 132 /

4 Table 1 Hot wall heat flux, pressure drop, and overall heat transfer efficiency in circular porous pin fin channel with different grids =0.9, PPI=30, T in =293 K, T h =343 K, Pr=0.7, Re =2291 Total elements 94, , ,080 q/kw m p/pa /kw m 2 Pa pin fin channels, respectively. The Nusselt number Nu hav and friction factor f are defined as follows: Nu hav = q D T h T in + T out /2 k f ; f = p/l D 1/2 f u 2 in where D=2H is the hydraulic diameter of inlet. The Reynolds number Re is defined as follows: Re = u in D 8 f The Darcy Number Da and average Nusselt number Nu av in the middle section z=0.5w of each heater for model validations see Figs. 4 and 5 are defined as follows: Da = K H 2 ; Nu q H av = 9 T hav T in k f y=0, z=0.5w where K is the permeability of porous blocks, H is the channel height, q is the constant heat flux of each heater, k f is the thermal conductivity of fluid, T in is the temperature of inlet, and T hav is the average temperature in the middle section y=0, z=0.5w of each heater. The governing equations Eqs. 1 3 for the computational domain are solved with commercial code CFX10. The convective term in momentum equations is discretized with high resolution scheme. The continuity and momentum equations are solved together with coupled solver based on finite control volume method and the discrete equations are solved with multigrid accelerated incomplete lower upper factorization technique CFX The user-define expressions for the additional energy equation of porous matrix T s and source terms of interphase heat transfer in both energy equations of fluid and porous matrix T f and T s, Eq. 3 are developed and compiled with CFX expression language. Furthermore, the conservative interface flux conditions for mass, momentum, and heat transfer are adopted at the interfaces between clear fluid and porous regions. For convergence criteria, the relative variations in temperature and velocity between two successive iterations are demanded to be smaller than the previously specified accuracy levels of Grid Independence Test and Model Validation Before proceeding further, the grids used for present study are checked at first. As shown in Fig. 2 a, the circular porous pin fin model is selected for the test and the computational parameters are kept constant with =0.9, PPI=30, T in =293 K, T h =343 K, Pr =0.7, and Re=2291 Air: u in =2 m s 1. In present test, a multiblock, O-type, structural grid with hexahedral elements is used and the grid is intensified on solid walls and pin fin regions. The total numbers of grid elements vary from 94,809 to 766,080 and the computational results are presented in Table 1. It shows that, the grid with total element number of 297,864 is good enough for the test case with the maximal lengths of the grid elements of being 0.47 mm for central flow region and 0.1 mm for near wall flow region. Therefore, similar grids are finally employed for the following studies and the total numbers of grid elements for different pin fin models are listed in Table 2. 7 Table 2 Computational grids for different pin fin models Pin fin models Circular Cubic Long elliptic Short elliptic Total elements air 297, , , ,512 Total elements water 147, , , ,120 Furthermore, the reliability and accuracy of present computational models and method are validated. According to the authors knowledge, most studies of forced convection heat transfer in partially filled porous channels were based on two-dimensional 2D model and almost no three-dimensional researches have been reported on this subject before. Therefore, a two-dimensional similar problem as reported by Hadim 5 see Fig. 3 a is finally selected for the validations. In the present study, the 2D partially filled porous channel 5 is extended along z coordinate with width of W=10H and a reasonable 3D physical model is finally obtained for the computation see Fig. 3 b. The 3D partially filled porous channel with dimensions of L H W is equipped with four porous blocks and each block is heated at bottom with constant heat flux. The inlet temperature and velocity are kept constant and all other walls are kept adiabatic. The computational model and method used for this problem are similar to those presented in Sec. 2 and the predicted average Nusselt numbers Nu av in the middle sections y=0, z=0.5w of different heaters are compared with those as reported in Ref. 5 see Fig. 4. The average deviation of Nu av is 3.5%. This indicates that the computational models and method presented in the present study are reliable and capable of modeling flow and heat transfer phenomena in 3D partially filled porous channels. 4 Results and Discussion 4.1 Performance Comparison for Solid and Porous Pin Fin Models. First, the flow and heat transfer performances in solid and porous pin fin channels are compared. The circular pin fin form see Fig. 2 a is selected for present study. Air Pr =0.7 and water Pr=3.9 are used as cold fluids and the Reynolds number Re varies from 1000 to 2291 with =0.9, PPI=30, T in =293 K, and T h =343 K. The temperature distributions in solid and porous pin fin channels are shown in Fig. 5. It shows that the internal temperatures of Fig. 3 Physical models for model validation: a 2D physical model reported in Ref. 5 and b 3D physical model used for present computation based on a / Vol. 132, MAY 2010 Transactions of the ASME

5 Fig. 4 Comparison of average Nusselt number of each heater with Ref. 5 solid pin fins are quite uniform and the average temperatures are high, which are K for air case and K for water case, respectively, while the internal temperatures of porous pin fins are not so uniform and the average temperatures are much lower, which are K for air case and K for water case, respectively. However, the fluid temperatures in porous pin fin channels are higher than those in solid pin fin channels. The average fluid temperatures in porous pin fin channels are K for air and K for water while they are K and K in solid pin fin channels. These results indicate that more heats can be transported away by using porous pin fins and their heat transfer performances would be better. This is because the porous pin fins can greatly enlarge the contact surface areas and mix the fluid flow inside, which may lead to significant heat transfer enhancements. The velocity vector distributions in solid and porous pin fin channels are presented in Fig. 6. It shows that with the same Reynolds number, the fluid velocities in solid pin fin channels are much higher than those in porous pin fin channels for both air and water cases. Large vortices are formed behind solid pin fins while no such vortices are found in porous pin fin channels. In solid pin fin channels, the solid pin fins are totally impermeable and this would narrow the flow passages and enhance the flow tortuosities inside. While in porous pin fin channels, the porous pin fins are permeable and the fluid can flow through them directly. This would widen the flow passages and lower the flow tortuosities inside. Therefore, the flow resistances and pressure drops in porous pin fin channels would be lower. The variations in pressure drop p, hot wall heat flux q, overall heat transfer efficiency, and heat transfer performance ratio with Reynolds number are presented in Fig. 7. It shows that the pressure drops in porous pin fin channels are much lower Fig. 6 Velocity vector distributions in solid and porous pin fin channels =0.9, PPI=30, Re=1000 : a solid pin fin channel air: y=5 mm, b porous pin fin channel air: y=5 mm, c solid pin fin channel water: y=0.5 mm, and d porous pin fin channel water: y=0.5 mm than those in solid pin fin channels 36.9% lower for air and 9.5% lower for water at Re=2291, see Fig. 7 a while the heat fluxes in porous pin fin channels are much higher than those in solid pin fin channels 38.6% higher for air and 45.7% higher for water at Re=2291, see Fig. 7 a. Therefore, the overall heat transfer efficiencies in porous pin fin channels are much higher 119.5% higher for air and 37.9% higher for water at Re=2291, see Fig. 7 b. It is also obvious that as Re increases from 1000 to 2291, all Fig. 5 Temperature distributions in solid and circular porous pin fin channels =0.9, PPI=30, Re=1000 : a solid pin fin channel with air, b porous pin fin channel with air, c solid pin fin channel with water, and d porous pin fin channel with water Fig. 7 Variations in pressure drop, hot wall heat flux, overall heat transfer efficiency, and heat transfer performance ratio with Re in solid and porous pin fin channels =0.9, PPI=30 : a pressure drop and hot wall heat flux and b overall heat transfer efficiency and heat transfer performance ratio Journal of Heat Transfer MAY 2010, Vol. 132 /

6 Table 3 Characteristics of metal foams Sample PPI d p /m d f /m K/m 2 c f k s /Wm 1 K the values of heat transfer performance ratios are larger than unit for both air and water cases see Fig. 7 b, which means that with the same pumping powers, the heat transfer performances in porous pin fin channels are also much better than those in solid pin fin channels. These results are consistent with the former analysis of temperature and flow variations, which confirms the point that with proper selection of physical parameters, the heat transfer augmentations and flow resistance reductions can be achieved simultaneously and the overall heat transfer performances will be significantly improved by using porous pin fins. Furthermore, it can be found that with different fluids, the flow and heat transfer performances are different. With the same Reynolds number, the pressure drops and heat fluxes in porous pin fin channels for water are much higher than those for air while the overall heat transfer efficiencies and heat transfer performance ratios are much lower. Due to the intrinsic thermophysical differences between air and water, the viscosity and heat transfer capacity of water are much higher, which would lead to higher pressure drops and heat fluxes. However, due to the same reasons, when water is used as cold fluid, most heats will be transported away just through the lower parts of the porous pin fins and the heat transfers in the upper parts of the porous pin fins are inactive see Fig. 5 d. Therefore, the overall utilization ratios of the porous fins are low for water, which would lead to lower overall heat transfer efficiencies and heat transfer performance ratios. 4.2 The Effect of Pore Density. In this section, the effect of pore density PPI for different metal foams is investigated. The circular pin fin form see Fig. 2 a is selected again for present computations. Both air Pr=0.7 and water Pr=3.9 are used as cold fluids. In the present study, five different kinds of high porosity metal foams with 20 PPI 40 are selected for the computations, which are similar to those as studied by Bhattacharya et al. 17 and Calmidi and Mahajan 18. These metal foams would also be common in industry applications and their characteristics are presented in Table 3. Furthermore, besides PPI, the other parameters are kept at constant with =0.9, T in =293 K, T h =343 K, and Re=2291. The variations in pressure drop p, hot wall heat flux q, and overall heat transfer efficiency with pore density PPI are presented in Fig. 8. It shows that, as PPI increases from 20 to 40, the pressure drops and heat fluxes in porous pin fin channels increase for both air and water cases while the overall heat transfer efficiencies decrease. This is because, as PPI increases, the permeability K decreases rapidly and the viscosity effects inside porous media increase, which would lead to increases in pressure drops. Meanwhile, as PPI increases, the solid-fluid interfacial surface areas inside porous media also increase quickly and the volumetric heat transfer coefficient h v between porous matrix and fluid phase increases, which would lead to increases in heat fluxes. As PPI increases from 20 to 40, the pressure drops in the porous pin fin channels increase by 52.9% and 24.2% for air and water, respectively, and the corresponding heat fluxes increase by 15.6% and 21.1%. It is obvious that the increase rates of pressure drops are much higher than those of heat fluxes, especially when air is used as cold fluid. Therefore, as PPI increases, the overall heat transfer efficiencies decrease. Furthermore, with different pore densities 20 PPI 40, the pressure drops in porous pin fin channels are lower than those in solid pin fin channels 50.5% lower for air and 20.2% lower for water at PPI=20 while the heat fluxes in porous pin fin channels are higher 23.4% higher for air and 30.1% higher for water at PPI=20. Therefore, the overall heat transfer efficiencies in porous pin fin channels are much higher and their maximal values are obtained at PPI=20, which are 149.2% and 63.1% higher than those in solid pin fin channels for air and water cases, respectively. These results indicate that with proper selection of pore density, the flow and heat transfer performances in porous pin fin channels will be improved. 4.3 The Effect of Porous Pin Fin Form. Finally, the effect of porous pin fin form is examined. Four different kinds of porous pin fins are compared here, including circular, cubic, long elliptic, and short elliptic cross-section forms see Fig. 2. Air Pr=0.7 and water Pr=3.9 are used as cold fluids and the Reynolds number Re varies from 1000 to 2291 with =0.9, PPI=40, T in =293 K, and T h =343 K. The temperature distributions in different porous pin fin channels are presented in Figs. 9 and 10. It shows that, the temperature distributions in circular and cubic porous pin fin channels are similar while they are quite different in long elliptic and short elliptic porous pin fin channels for both air and water cases. The average temperatures of air are K, K, K, and K in circular, cubic, long elliptic, and short elliptic porous pin fin channels, respectively, and they are K, K, K, and K for water, respectively. It is obvious that, the average temperatures of air and water in short elliptic porous pin fin channels are the highest and they are the lowest in long elliptic porous pin fin channels. The flows in short elliptic porous pin fin channels are intensively mixed and most of the fluids in the channels, including central flow regions, have taken part in heat transfer actively while in long elliptic porous pin fin channels, the Fig. 8 Variations in pressure drop, hot wall heat flux, and overall heat transfer efficiency with pore density =0.9, Re= / Vol. 132, MAY 2010 Transactions of the ASME

7 Fig. 9 Temperature distributions in different porous pin fin channels with air =0.9, PPI=40, Pr=0.7, Re=1000 : a circular pin fin channel, b cubic pin fin channel, c long elliptic pin fin channel, and d short elliptic pin fin channel Fig. 11 Variations in pressure drop, hot wall heat flux and overall heat transfer efficiency with Re in different porous pin fin channels =0.9, PPI=40, Pr=0.7 Fig. 10 Temperature distributions in different porous pin fin channels with water =0.9, PPI=40, Pr=3.9, Re=1000 a circular pin fin channel, b cubic pin fin channel, c long elliptic pin fin channel, and d short elliptic pin fin channel flows are less mixed, the central flow regions are almost not disturbed. Therefore, the heat transfers in short elliptic porous pin fin channels would be the highest. However, with the same reasons, the pressure drops in short elliptic porous pin fin channels would also be the highest. The variations in pressure drop p, hot wall heat flux q, and overall heat transfer efficiency with Reynolds number are presented in Figs. 11 and 12. It shows that with the same Reynolds number, the pressure drops and heat fluxes in short elliptic porous pin fin channels are the highest and they are the lowest in long elliptic porous pin fin channels. The differences in pressure drops between each other are 95.9% and 48.7% for air and water cases at Re=2291, respectively, and the corresponding differences in heat fluxes are 60.6% and 14.3%. However, the variations in overall heat transfer efficiencies are reverse, which are the highest in long elliptic porous pin fin channels and lowest in short elliptic porous pin fin channels and the differences between each other are 21.9% and 30.1% for air and water cases at Re=2291, respectively. These results indicate that, with proper selection of porous pin fin forms, the overall heat transfer performances in porous pin fin channels will be greatly improved and optimized. 5 Conclusions The forced convective heat transfer in three-dimensional porous pin fin channels is numerically studied in this paper. Both air and water are used as the cold fluids and the effects of Reynolds number, pore density, and pin fin form are carefully investigated. Fig. 12 Variations in pressure drop, hot wall heat flux and overall heat transfer efficiency with Re in different porous pin fin channels =0.9, PPI=40, Pr=3.9 The flow and heat transfer performances in porous pin fin channels are also compared with those in traditional solid pin fin channels in detail. The major findings are as follows. 1 With proper selection of metal foams, such as PPI=30, significant heat transfer enhancements and pressure drop reductions can be achieved simultaneously by using porous pin fins for both air and water cases, and the overall heat transfer efficiencies in porous pin fin channels are much higher than those in solid pin fin channels, which are 119.5% and 37.9% higher for air and water cases at Re =2291, respectively. Journal of Heat Transfer MAY 2010, Vol. 132 /

8 2 The effects of pore density are significant. As pore density increases from 20 to 40, the maximal overall heat transfer efficiencies are obtained at PPI=20 for both air and water cases, which are 149.2% and 63.1% higher than those in solid pin fin channels at Re=2291, respectively. 3 The effects of pin fin form are also remarkable. With same physical parameters, the pressure drops and heat fluxes are the highest in short elliptic porous pin fin channels and lowest in long elliptic porous pin fin channels while the overall heat transfer performances are the highest in long elliptic porous pin fin channels and lowest in short elliptic porous pin fin channels. The differences in overall heat transfer efficiencies between each other are 21.9% for air case and 30.1% for water case at PPI=40 and Re=2291, respectively. Acknowledgment We would like to acknowledge financial support for this work provided by the National Natural Science Foundation of China Grant No Nomenclature A area m 2 c F Forchheimer coefficient, Eq. 4 c p specific heat at constant pressure J kg 1 K 1 D hydraulic diameter of inlet m Da Darcy number, Eq. 9 d f fiber diameter of metal foam m d p pore size of metal foam m f friction factor, Eq. 7 G shape function for metal foam, Eq. 5 H channel height m h v volumetric heat transfer coefficient, Eq. 4 W m 3 K 1 K permeability, Eq. 4 m 2 k thermal conductivity W m 1 K 1 L total channel length m L 1 length of developing inlet block m L 2 length of hot wall m L 3 length of developing outlet block m Nu av average Nusselt number, Eq. 9 Nu hav average Nusselt number of hot wall, Eq. 8 p pressure Pa PPI pore density Pr Prandtl number q heat flux of hot wall, Eq. 6 Wm 2 q constant heat flux of each heater, Eq. 9 Re Reynolds number, Eq. 8 T temperature K u,v,w velocity in x, y, z directions m s 1 V velocity vector m s 1 W channel width m x,y,z coordinate directions m Greek Symbols overall heat transfer efficiency, Eq. 6 W m 2 Pa 1 heat transfer performance ratio, Eq. 6 kinetic viscosity m 2 s 1 density kg m 3 porosity tortuosity of porous matrix, Eq. 5 Subscripts f fluid phase, fiber h hot wall hav average value of each heater in inlet out outlet p value obtained in porous pin fin channel pin pin fin cross-section s solid phase, value obtained in solid pin fin channel References 1 Sahiti, N., Lemouedda, A., Stojkovic, D., Durst, F., and Franz, E., 2006, Performance Comparison of Pin Fin In-Duct Flow Arrays With Various Pin Cross- Sections, Appl. Therm. Eng., 26, pp Sahiti, N., Durst, F., and Dewan, A., 2005, Heat Transfer Enhancement by Pin Elements, Int. J. Heat Mass Transfer, 48, pp Sahiti, N., Durst, F., and Dewan, A., 2006, Strategy for Selection of Elements for Heat Transfer Enhancement, Int. J. Heat Mass Transfer, 49, pp Jiang, P. X., Li, M., Lu, T. J., Yu, L., and Ren, Z. P., 2004, Experimental Research on Convection Heat Transfer in Sintered Porous Plate Channels, Int. J. Heat Mass Transfer, 47, pp Hadim, A., 1994, Forced Convection in a Porous Channel With Localized Heat Sources, ASME J. Heat Transfer, 116, pp Hadim, A., and Bethancourt, A., 1995, Numerical Study of Forced Convection in a Partially Porous Channel With Discrete Heat Sources, ASME J. Electron. Packag., 117, pp Huang, P. C., and Vafai, K., 1994, Analysis of Forced Convection Enhancement in a Channel Using Porous Blocks, J. Thermophys. Heat Transfer, 8, pp Huang, P. C., Yang, C. F., Hwang, J. J., and Chui, M. T., 2005, Enhancement of Forced-Convection Cooling of Multiple Heated Blocks in a Channel Using Porous Covers, Int. J. Heat Mass Transfer, 48, pp Fu, W., Huang, H., and Lion, W., 1996, Thermal Enhancement in Laminar Channel Flow With a Porous Block, Int. J. Heat Mass Transfer, 39, pp Ould-Amer, Y., Chikh, S., Bouhadef, K., and Lauriat, G., 1998, Forced Convection Cooling Enhancement by Use of Porous Materials, Int. J. Heat Fluid Flow, 19, pp Ko, K. H., and Anand, N. K., 2003, Use of Porous Baffles to Enhance Heat Transfer in a Rectangular Channel, Int. J. Heat Mass Transfer, 46, pp Yang, Y. T., and Hwang, C. Z., 2003, Calculation of Turbulent Flow and Heat Transfer in a Porous-Baffled Channel, Int. J. Heat Mass Transfer, 46, pp Yang, J., Zeng, M., and Wang, Q. W., 2008, Numerical Study on Forced Convective Heat Transfer in Porous Pin Fin Channels, ASME Paper No. HT Ergun, S., 1952, Fluid Flow Through Packed Columns, Chem. Eng. Prog., 48, pp Wakao, N., and Kaguei, S., 1982, Heat and Mass Transfer in Packed Beds, McGraw-Hill, New York. 16 Nield, D. A., and Bejan, A., 2006, Convection in Porous Media, 3rd ed., Springer, New York. 17 Bhattacharya, A., Calmidi, V. V., and Mahajan, R. L., 2002, Thermophysical Properties of High Porosity Metal Foams, Int. J. Heat Mass Transfer, 45, pp Calmidi, V. V., and Mahajan, R. L., 2000, Forced Convection in High Porosity Metal Foams, ASME J. Heat Transfer, 122, pp CFX Inc., 2005, User Guide, CFX / Vol. 132, MAY 2010 Transactions of the ASME